Seven Shock-transformed Hexahedrites, Sensitive to Grainboundary Corrosion V. F. Buchwald Department of Metallurgy, Technical University, Lyngby, Denmark

Z. Naturforsch. 35a, 361-369 (1980); received January 21, 1980*

Dedicated to Professor II. Hintenberger on the occasion of his 70th birthday.

Seven polycrystalline (Forsyth County, Holland's Store, Wathena, Pima Co., Chico Mountains, Kopjes Vlei and Mejillones), of bulk composition as normal hexahedrites of group 11 A, are shown to be severely altered hexahedrites. Mic-roprobe examination and various heat treatments of synthetic alloys support the conclusion that the meteorites after a primary slow cooling were subjected to shock-reheating above 800°C. Thereby the following transforma- tions took place a-^-y—>a2—>a, and an unequilibrated structure, very sensitive to grain-boundary corrosion was formed.

Introduction Table 1. Weight and bulk composition of the seven polv- crystalline iron meteorites. Other important elements are The last ten years have seen a very significant iron (about 93%), cobalt (0.5%) and sulfur (0.1%). progress with respect to understanding the structure Weight percent PPm and composition of iron meteorites. The efforts in analysing for main and trace elements have resulted kg Ni P Ga Ge Ir in a classification system [1, 2] that may form the Forsyth basis for all further discussions. Mass-spectrometric County 23 5.54 0.21 60 176 31 work [3, 4] has provided valuable insight in the Holland's Store 12.5 5.35 0.25 61 184 20 isotopic composition and the content of noble gases, Wathena 0.57 5.54 0.27 60 184 7.0 data which allow us among other things to deduce Pima the age of meteorites, and the intensity of the cosmic County 0.21 5.60 0.25 60 181 8.9 Chico' radiation. Mountains 0.21 5.56 0.33 59 176 6.2 However, certain meteorites have for various Kopjes Vlei abt. 8 5.65 0.3 59 181 3.1 reasons been difficult to handle. One such group Mejillones abt. 15 5.67 0.33 60 177 2.3 consists of seven polycrystalline iron meteorites, The analyses are weighted averages as reported in [2], Table 1, which because of their enigmatic structure, that sets them apart from the more familiar hexa- hedrite or Widmanstätten structures, only rarely long puzzled scientists, are mainly artefacts, and that have been included in systematic investigations. the extreme sensibility to terrestrial corrosion is due Apart from brief descriptions with photomicrographs to a high degree of unequilibrium, which has sensi- [5, 6] and more extended ones by Buchwald [2], tized the grain boundaries. Finally, an attempt will the meteorites in question have only been cursorily be made to present an evolutionary path that might examined [7 — 9]. In the Hey-Catalogue [10], they lead to the observed structures. are indiscriminately labelled -poor , brecciated hexahedrites or granular metabolites. Structure The purpose of this paper is to explain the struc- ture and by experiments to support the presentation. While the seven meteorites do have some structural It will be shown that the microscopic holes, that have characteristics that may serve to distinguish them from each other, the following presentation will emphasize what they have in common: The poly- Reprint requests to Dr. V. F. Buchwald, Department of crystalline , the shock-melted ag- Metallurgy, Technical University, Building 204. 2800 Lyng- by, Dänemark. gregates, the resorbed phosphides, and the general * The first version received October 4, 1979. unequilibration of all structural elements.

0340-4811 I 80 / 377-0370 S 01.00/0. — Please order a reprint rather than making your own copy. 362 V. F. Buchwald • Shock-transformed Hexahedrites

Polished sections show a massive material in , which once rimmed the troilite nodule, which are embedded scattered diffuse patches of Figure 4. troilite and minute particles of graphite. Graphite occurs as angular or spheroidal, often Upon etching a grain structure, which is unique zoned aggregates, 10 — 150 um across. The spherules to these iron meteorites, develops. Almost equiaxial are composed of anisotropic, radiating crystallites grains, 0.3 — 3 mm in diameter, constitute the bulk and are most common near the schreibersite rem- of the specimens. Each kamacite grain is differently nants, presumably because the graphite was derived oriented as deduced from the oriented sheen from from , previously rimming the schreibersite. etching, and from the presence of occasional Neu- Figure 5. mann bands. Two of the meteorites, Mejillones and Kopjes

Each grain is divided into many thousands almost Vlei, are no more corroded, than the heat-affected a2 equiaxial subgrains, 30 — 50 um in diameter, Fig- zone and an occasional patch of the fusion crust have ures 1 — 2. Within each main-grain the subgrains are been preserved. The others are rather heavily cor- uniformly oriented, since Neumann bands pass the roded. Cohen [11, 12] and others after him, believ- subgrain boundaries with shifts of directions that ed that the strong corrosive attack was due to lawren- are less than 1 . cite, FeClo , a mineral which was assumed to be part The hardness of the metalphase is 190 + 20 of the structure. As discussed previously (\ickers DPN, 100 g load), somewhat dependent on [13] the chlorine is, however, not of cosmic origin, the local cold work from late splitting in the atmo- but has entered the grain boundaries during a long sphere in connection Avith the fall. exposure to terrestrial groundwater. Grainboundaries and subboundaries are all very Pima County and Chico Mountains, which are dirty, rich in tiny precipitates, 0.2 — 2 um across, verv small and of a fragmental nature, have an ob- which on the microprobe could be shown to be scure origin and may very well have come from the mainly iron-nickel-phosphides. It is, however, pos- same, now lost meteorite [2, p. 976]. sible that some of the smaller ones are carbides. Forsyth County and Holland's Store have, in ad- Equally distributed throughout the sections with dition to the structures discussed above, at one end a frequency of 300 - 500 pr. mm2 there occurs a a different, more hexahedrite-like structure. The two puzzling unit, a rounded or elongated metal grain structural types are separated by a wavy boundary. 3 — 15 um across, normally associated with liny Forsyth County and Holland's Store are in all phosphide particles. These "unequilibrated loops" details so similar that they, too, might come from may be situated on grainboundaries or away from ihe same original fall. them, Figure 3. There are no rhabdite-prisms or -plates in these The problematic pits meteorites. There is, however, occasional large schreibersite crystals, up to 10 X 1 mm in size. These The structural description above was based upon are severely damaged and partially dissolved. the study of perfect sections. However, upon routine Around them are zones with steep gradients of polishing and etching the sections always become nickel and phosphorus, and ihe structure shows a pitted, wherefore most published photomicrographs corresponding zonation in grain size, and frequency (e. g. [5, 6]) show a profusion of black dots. There and type of the tiny precipitates, Figure 8. Nearest has been considerable confusion as to the cause of the phosphide are equal numbers of amoeboid the pits; it has thus been proposed that they rep- particles and angular phosphides, farther away only resented oxides, magnetite [5], altered rhabdites phosphides occur. [2], or were pores in non-massive meteorites [7]. Cm-sized troilite-daubreelite aggregates of the The first result of this investigation is that the usual lamellar development, as seen in e. g. pits are artifacts. They are not genuine and the Coahuila, were once present. Now they have been meteorites are not porous. Careful cutting and transformed into intricate intergrowths of 1 — 10 um polishing, supported by impregnation, and perform- grains of ironsulfide, iron-nickel-sulfide and iron, ed with small load and good cooling and lubrication, with dispersed fragments of daubreelite, and of will usually result in perfect samples with no pitting. Fig. 1. Forsyth County. Corrosion penetrates from the surface (left) along Fig. 2. Forsyth County. The uncorroded interior displays a polycrystalline the grain boundaries. The unequilibrated loops are selectively attacked kamacite structure with a significant number of sensitive, unequilibrated and the polished surface becomes pitted (black). Etched. Scale bar 40 /mi. loops. Lightly etched. Scale bar 40 //m.

Fig. 4. Wathena. Detail of a shock-melted troilite nodule. The rapid melting Fig. 3. Forsyth County. Four unequilibrated loops showing various stages and solidification produced a fine-grained eutectic of iron and ironsulfide of destruction, caused by polishing and etching. To the right several in- with dispersed daubreolito (dark grcv) and schreibersite fragments (white), distinct Neumann bands. Oil immersion. Scale bar 10/mu. hitched. Oil immersion. Scale bar 10//in. 364 V. F. Buchwald • Shock-transformed Hexahedrites

However, since the pits develop so easily in these are soft, about 145 — 150 Vickers, except when they and not in others of the same bulk composition, have been cold-worked by fragmentation in the some explanation must be sought. atmosphere. The perfect polished section is massive and it is The table shows that phosphorus is increased by hardly possible to predict that pits will occur during about a factor five in the pitted meteorites. Simul- subsequent handling. Already a very weak etching taneously the volume-percent of phosphides is signifi- with 1% Nital {1% nitric acid in ethanol) will cantly decreased. Clearly, the kamacite of the poly- reveal the grain- and subgrain boundaries, and their crystalline meteorites is supersaturated with respect associated precipitates. A somewhat longer etching to phosphorus. deepens the grain boundaries, and especially the There are indications that the metallic phases are "unequilibrated loops'' are severely attacked. Final- supersaturated with respect to carbon and sulfur ly, the loops become undercut and 1 — 15 lira large atoms too, but this assumption could not be verified grains easily pop out, leaving the characteristic with the microprobe. However, in at least one of the rounded pits. The phosphide particles usually remain seven meteorites, Kopjes Vlei, carbon is clearly con- in situ. After prolonged etching, the pits may assume centrated in the unequilibrated loops. The carbon

angular outlines, corresponding to the crystallo- reveals itself in the heat affected a.2 zone where the graphic orientation of the kamacite grains. It ap- loops now are surrounded by darketching nickel- pears that there is a 1 : 1 correlation between the bainite zones, created by diffusion of carbon away unequilibrated loops and the pits. from the loops during the brief atmospheric reheat- Microprobe work only revealed the usual com- ing, Figure 6. ponents, kamacite and phosphide, plus occasional Optical examination, hardness testing and micro- graphite and sulfide. The phosphides have a fre- probe work thus all lead to the same conclusion, that quency and distribution reminiscent of rhabdites in the seven meteorites are unequilibrated and in- hexahedrites. They are, however, much smaller than homogeneous, both in structural, mechanical and in hexahedrites and usually occur as angular, ir- chemical respects. It is this lack of equilibration that regular blebs in a — a grain boundaries. leads to the pitting. The chemical gradients around Results from microprobe work on three of the the grain boundaries, possibly aided by ultrathin seven meteorites are shown in Table 2; work on graphite films on some boundaries and by mechanical Boguslavka and North Chile has been included for stress-concentration, concentrate the attack along the comparison. These two are hexahedrites of the same sensitized boundaries and loosen the coherence. Pits bulk composition, but monocrystalline and presum- develop where the small grains become undercut, ably representing a much better state of equilibrium otherwise only heavy grain boundary etching devel- because they have a very low cooling rate, on the ops. order of 1 per million year [14] (this reference During exposure to terrestrial ground water a discussed Quillagua and Tocopilla, which both are similar attack, albeit more slowly, develops. Chloride- fragments of North Chile). The normal hexahedrites containing waters are particularly active and pene-

Table 2. Three of the polycrystalline meteorites compared to two normal hexahedrites, Boguslavka and North Chile.

Bulk composition11 Kamacite composition1' Phosphide Reference °0Xi ppm P °0Ni ppm P typec

Forsvth Countv 5.54 2100 5.4 2010 4- 220 GBP 9 Forsvth Countv 5.54 2100 5.3 ± 0.2 2200 4- 500 GBP this work Kopjes Vlei 5.65 3000 5.4 + 0.3 2000 — 500 GBP this work Mejillones 5.67 3300 5.81 ± 0.4 2620 ± 300 GBP S Mejillones 5.(57 3300 5.72 ± 1.2 2850 - 700 GBP 8 Boguslavka 5.42 2000 5.2 ± 0.1 500 + 100 Pv this work North Chile. Puripica 5.59 3000 5.28 ± 0.2 400 4- 100 Pv 8 North Chile. Yungav 5.59 3000 5.30 ± 0.1 500 ± 100 Pv this work

a Data as compiled in [2]; '» Microprobe data; c GBP, mainly grain boundary precipitates. R mainly rhabdites. 365 V. F. Buchwald • Shock-transformed Hexahedrites

träte from the surface along the grain boundaries. The resulting structure was in all essential features The small loops are most readily converted to identical to that of the meteorites: Polycrystalline limonite, later follows the bulk of the material ([2], ferrite with fine phosphide precipitates and a dense Figures 1024, 1905). population of unequilibrated loops, Figure 7. The heat treatment is an example of the solid- state transformations Experimental Verification heating cooling cooling In order to establish a better understanding of the formation of the unequilibrated structure, extensive 800" ' 650° "600-450° experiments on synthetic alloys have been perform- On cooling, the polycrystalline y first transforms ed. In this connection only a few pertinent data will diffusionless to a2 . Then, on further slow cooling, be discussed. the a2 grains proceed towards a polycrystalline, The alloys (Table 3) were produced from powders equiaxed a structure of improved equilibrium. The of carbonyl iron, carbonyl nickel and iron phosphide impurity atoms, P and C, interfere with the grain [15]. boundary movements and prevent, under these cool- Compression, followed by sintering for at least ing rates of about 25° per day, the formation of 24 hours in hydrogen at 1100 °C, and ending with perfect grains. Instead a large number of unequili- quenching in water, resulted in alloys, that under brated loops are left pinned in the structure, as seen microprobe examination proved to be homogeneous in Fig. 7, which should be compared to Fig. 2 of the within ± 0.05% nickel and phosphorus. meteorites. The alloys contained as an impurity 1000 ppm Si max. While this is about a factor 50 more than is A) Rapid heating followed by rapid cooling usual for iron meteorites [2, 16], the concentration Samples were then reheated above the a —> y of silicon is still too low to have had any significant transformation temperature, which for these alloys effect on the equilibrium diagrams here in question. is 750 — 850°. They were held for half an hour and However, the silicon atoms, which are substitutional again quenched in water. The resulting structure was in the Fe-Ni-lattice, may be assumed to have rein-

the well-known a2 structure, a rather hard, unequi- forced the pinning of the P and S atoms. librated structure, which is typical for the heataffect-

ed a2 zone on fresh falls of hexahedrites. It did, Discussion however, not correspond to the structure of the polycrystalline meteorites in question. The seven enigmatic meteorites have bulk com- positions and trace element concentrations, which within the limits of analytical error place them in B) Rapid heating followed by slow cooling group II A [1, 2]. They thus constitute a significant Alloys of the same composition were reheated to subset of the total population of about 45 meteorites similar high temperatures in the 7-region, and then in group II A. left to controlled, slow cooling. In a typical ex- It is generally assumed that the hexahedrites have periment the cooling from 900 to 450° lasted 20 had a very small cooling rate whereby they develop- days, whereafter the furnace was held at 450 for an ed a coarse kamacite structure with grain-sizes of at additional 5 days. Then the batch was quenched. least 30 cm diameter [2, 14]. It might be postulated therefore that the fine-grained meteorites here discus- sed represent material which had been subjected to Table 3. Composition of synthetic alloys, weight % a much more rapid primary cooling. This appears, however, not to be the case, since 0 Xi 0/ p 0/ N 0/ Si 0 ^1 /o r 0 v 0 %Cu remnants of troilite and schreibersite are present in

E 3 3.0 0.5 0.03 0.10 0.02 a morphology and size which are typical of hexa- G 46 4.0 0.6 0.03 0.10 0.02 hedrites. The centimetersized inclusions of schreiber- G 57 5.0 0.7 0.03 0.10 0.02 site must have been precipitated by very slow solid- E 6 6.0 0.5 0.03 0.10 0.02 state reactions below about 600 °C [17]. Therefore, Fig. (i. Kopjes Vlei. Detail of the heat-affected ao rim zone from atmospheric Fig. 5. Mejillones. Polycrystalline kamacite with numerous angular phos- flight. The brief reheating to above 1000°C here melted and rounded the phides in the deep-etched grain boundaries. Below, a zoned graphite pre- phosphides (P), and carbon diffused away from the unequilibrated loop cipitate. Etched. Scale bar 25 //m. and formed darketching nickel-bainite. Etched. Oil immersion. Scale bar 20 inn.

r> ^ o e \ ' \ . r • Ö * •Q •

' 0 Ox . ,

>o P. ° . . o v • •» ~ 19kt i

O

Fig. 7. E .'5. subjected to 20 days of slow cooling from 900 to 450°0. Fig. S. Mejillones. Around a primary schreibersite crystal there is a strongly Polycrystalline ferrite with a number of unequilibrated loops, which upon zoned variation in the phosphorus and nickel concentration, apparently further polishing and etching will develop into pits. Slightly etched. Scale caused by a brief, secondary, non-equilibrium thermal cycle. Etched. bar 20 /mi. Scale bar 1 mm. 367 V. F. Buchwald • Shock-transformed Hexahedrites the fine-grained meteorites must at an earlier time asteroidal-sized body, but rather a meter-sized frag- have been coarse-grained. ment. In other words, the shock that produced the In the following model an attempt is made to in- structural alterations, also fragmented the parent corporate all relevant structural and compositional body and released the meteorites which from now observations. A primary slow cooling generated on were exposed to cosmic radiation. typical, coarse-grained hexahedrite material with The only later event, which there is evidence for occasional nodules of cm-sized troilite, and with in the meteorites, is the penetration of the atmo- mm-cm-sized schreibersite crystals. Thin cohenite sphere, during which the already small masses were rims developed around some of the crystals, and in further reduced and fragmented. The surface was the matrix a heavy population of micron-sized rhab- reheated and ablated, and Neumann bands formed dites precipitated. This structure is well-exemplified in grains along the intergranular cracks [2]. by, e. g., North Chile ([2], Figures 1288 - 1298). The stages of shock-metamorphism and hexa- A severe shock with elements of shearing reheated hedrites may be illustrated by Figure 9. After a the material to above 800 °C bulk temperature. In slow primary cooling the meteorites-to-be are cool, this phase, which probably lasted less than an hour, soft, well-annealed, coarse-grained and with equilib- the schreibersite and rhabdite in the metal phase rium precipitates, point A. The shock-event deforms started to dissolve, and cohenite decomposed. The and reheats the material to a degree which depends monocrystalline kamacite recrystallized to poly- on distance from the impact point and on the local crystalline y, which grew to mm-sized grains. The geometry and shielding. The distant part will be only pulse of heat sent waves of P-, Ni- and C-atoms slightly reheated, leading to recovery. Another part away from the preexisting inclusions, but equilib- will be reheated above 500 °C and thereby re- rium was not achieved before cooling set in again, crystallize provided previous cold-work were suf- Figure 8. ficient. Finally, a small part will be reheated above Even if the bulk temperature attained was only the a—>y transformation temperature and behave as about 800 °C, the compressible troilite inclusions discussed in the present paper. readied temperatures above 1000 °C, sufficient to If the meteorites of group II A are subjected to a melt them. The associated daubreelite and adjacent structural reevaluation, we see that we have samples schreibersite rims were fragmented and disseminated of all categories, Table 4. The total population may into the melt, Figure 4. therefore be taken to represent a which Very soon the material started to cool, with a at one time was subjected to a violent shock, and was cooling rate about 25° per day. Thereby the split and transformed. Of course, some of the numerous ;'-crystals independently transformed to material melted and vaporized, but of this nothing OL.y, whereupon these slowly changed to more equi- has survived, or at least has not been identified as axed a-grains. Part of the phosphorus reprecipitated meteorites so far. as phosphide particles, while part of the carbon The situation has a certain similarity to the impact reprecipitated as graphite nodules and ultrathin fdms event at Barringer Crater, Arizona. Around this on some grain-boundaries. The cooling-rate was high crater a large number of metallic fragments, the enough to prevent equilibrium with respect to Canyon Diablo meteorites, are scattered. Metallo- phosphorus, wherefore about 0.2% P remained in graphie analysis ([2], page 390) shows that these solid solution. The impurity atoms and precipitates fragments represent all stages of shock-transforma- interfered strongly with the grain boundaries and tion and many of them display steep temperature resulted in a large number of unequilibrated loops. gradients. In addition, also the shock-melted rem- This shock-metamorphism will thus fit into our nants have survived and may be distinguished as tiny, general understanding of shock-processes, but adds metallic droplets scattered over the arid plain. a new dimension, in which a via y forms ct2, which The boundary, that in Holland's Store and Forsyth then approaches equilibrium a, of a peciliar ap- County separates coarse- and fine-grained material, pearance and of an easily corroded character. suggests that we also here have had steep temper- Since the material, after having reached the ature gradients. In the coarse-grained end the tem- maximum temperature, cooled rather rapidly it must perature did not exceed about 750 C, therefore the be assumed that it was no longer part of a large, a-structure did not transform to y which was the 384 V. F. Buchwald • Shock-transformed Hexahedrites

Cooling rate 25 per day

ci2 •* polycrystalline a

Shock-transformed hexahedrites

Recrystailized hexahedrites

Cosmic radiation age

Fig. 9. Sketch suggesting the various reaction paths resulting from shock-reheating. The thermal cycle leads to signi- ficantly different structures for the same bulk composition.

Table 4. Some hexahedrites of group IIA, arranged accord- Conclusions ing to the extent of secondary changes in the primary structure. i) The seven fine-grained meteorites belong to the hexahedrite class, chemical group II A. Structural alterations Examples ii) They cooled slowly together with the bulk of group II A and developed an annealed coarse-grain- Only small North Chile, Walker County, ed hexahedrite structure. structural changes Murphy, Smithonia iii) A shock event disrupted the parent body, and Increasing recovery, Bruno, Bennett County steep temperature- and pressure gradients developed. recrystallization Angra dos Reis and decomposition Scottsville, Edmonton The fine-grained meteorites represent a case of of minerals Boguslavka severe alteration, where the reaction path a (mono- Sierra Corda, Negrillos Cedartown crystalline)-^ y (polycrystalline)-*- ot2 —> a (polycrys- talline) was followed. a -> y transformation, Forsyth County, Holland's Store decomposed minerals, Wathena, Pima County, iv) The included minerals decomposed, and the and unequilibrated Chico Mountains, Kopjes Vlei cooling-rate, of the order of 25° per day, was so matrix Mejillones high that a significant proportion of the phosphorus remained in solid solution. v) The impurities of phosphorus and carbon necessary step for further decomposition to au. In prevented a full equilibration of grain boundaries. the opposite end, 20 cm away, the temperature was Tiny phosphides and carbides, and graphite segre- high enough for the transformations to take place. gates, line many of the grain boundaries. 369 V. F. Buchwald • Shock-transformed Hexahedrites

vi) Terrestrial corrosion attacks preferentially AcJcnowledgements along these unequilibrated grain boundaries. The difficult problems associated with polishing vii) Routine cutting and polishing develop a and etching the samples were largely solved due to large number of micron-sized pits, conditioned by the assistance of Mrs. Edith Johannsen. The as- the said lack of equilibration. sistance of R. Norbach Nielsen at the microprobe, viii) Lawrencite was never present as a cosmic and support from the National Foundation of Tech- mineral; the chlorine in the grain boundaries is nical Sciences (STVF grant 516.5011-K.288) is derived from terrestrial groundwater. gratefully acknowledged.

[1] J. T. Wasson, Meteorites-Classification and Proper- [10] Max H. Hey, Catalogue of Meteorites, British Museum, ties. Springer-Verlag, Berlin 1974. London 1966. [2] V. F. Buchwald, Handbook of Iron Meteorites. Uni- [11] E.Cohen, Meteoritenkunde, Heft II, Stuttgart 1903, versity of California Press. Vol. 1 — 3. 1975. 264. [3] H. Hintenberger, L. Schultz, H. Wänke, and H. [12] E. Cohen, Meteoritenkunde, Heft III, Stuttgart 1905, Weber, Z. Naturforsch. 22a, 780 (1967). 77. [4] H. Voshage and H. Feldmann, Z. Naturforsch. 35a, [13] V. F. Buchwald, Phil. Trans. Roy. Soc. London 50 [1980]. A286, 453 (1977). [5] S. H. Perry, The Metallography of Meteoritic Iron. [14] E. Randich, and J. I. Goldstein, Geochim. Cosmo- U.S. Natl. Museum, Bulletin 184. 1944. chim. Acta 42, 221 (1978). [6] A. A. Smales, D. Mapper, and K. F. Fouche, Geochim. [15] V. F. Buchwald, Acta Polytechnica Scandinavia, Cosmochim. Acta 31, 673 (1967). Ch 51 (1966), 46 pages. [7] H.J. Axon, Mineralogical Magazine 36, 1139 (1968). [16] C. M. Wai and J. T. Wasson, Geochim. Cosmochim. [8] J. T. Wasson and J. I. Goldstein, Geochim. Cosmo- Acta 33, 1465 (1969). chim. Acta 32, 329 (1968). [17] J. I. Goldstein and A. S. Doan, Geochim. Cosmochim. [9] S. J. B. Reed, Meteorite Research (editor P. M. Mill- Acta 36, 51 (1972). man) 1969, 743.